Equine infectious anemia virus (EIAV) is a retrovirus closely related to the primate immunodeficiency viruses, HIV and SIV. EIAV causes a persistent infection in horses, characterized by episodes of debilitating illness separated by asymptomatic periods. Degree of virulence (harm to the host) differs among forms of the virus; virulent forms of the virus rapidly evolve from avirulent forms when populations of virus are transmitted (passaged) from animal to animal. Recent work has identified specific regions of the genome that appear to be crucial in virulence determination, in particular the long terminal repeat (LTR) region and that encoding the envelope protein (env). Molecular clones are available for forms of the virus with known sequences for these regions and known virulence phenotypes, ranging from avirulent to highly virulent. In the proposed work, test animals will be infected with a molecular clone of an avirulent form of the virus and patterns of sequence diversification and divergence will be tracked through early stages of infection via intensive sequence analysis of the entire env/3' LTR region. Parallel serial passage experiments will be used to trace changes in levels and patterns of variation that occur with infection of a new host, with emphasis on identification of patterns of selection that result in predominance of a virulent form of the virus. This work will also test whether convergent changes occur at the molecular level when virulent forms evolve independently, but from the same starting avirulent clone. The extensive data collected should reveal a great deal about the evolution of virulence in general, and will allow description of the generation and standing levels of variation in virus within hosts as well as rate of divergence among hosts. An integrative approach to data analysis will be used, combining population genetic techniques with phylogenetic methodology to determine whether events early after infection, and the evolution of virulence that accompanies transmission, follow a predictable course with respect to pattern and process. This work should also yield insight into the molecular-level changes that are necessary for virulence to evolve. The EIAV system provides an excellent and tractable experimental model for study of evolutionary change in retroviruses, and for understanding the evolution and determinants of virulence.